Viscosity of liquid toluene at temperatures from 25 to 150.degree.C

Bank Development Project with the Institute of Technology Bandung, Indone- ... Institute for Physical Science and Technology, University of Maryland, ...
0 downloads 0 Views 1MB Size
J. Chem. EW. Data 1992, 37,349-355

Litoratwe Cited (1) Kwnvddsn, K. A. cvabelBkqsodwwn. Cy& 1088, 2, 221. (2) umuh,C. H.; Katz, D. L. Trans. A I M I O 4 0 , 168, 83. (3) Bemcz, E.; Rah-Achr, M. Research report No. 37 (185-XI-1-1974 OQIL); NMR, Tech. Unlv. Heavy Ind.: Wok,as summarked in a s m t a 9 ; Stu#ss h I r t t v g m k CHenntirby; Elsevier: New York, 1983; Vd. 4, 343. (4) ROMnm, D. B.; Mdrta, B. R. J . Can. Pet. T6dmol. 1071, 10, 33. (5) Adaamlto, 5.; Frank, R. J., 111; Skan, E. D., Jr. J . Chem. Eng. Date 1001, 38, 68. (6) baton, W. M.; Frost, E. M., Jr. U S . B u . Mhss 1948, 8. (7) Jhavd, J.; RoblMOn, D. B. Can. J . Chem. Eng. 1085, 43,75. (8) Holder, G. D.; bnd, J. H. AIChE J . 1082, 28, 44. (9) Vama, V. K. Gas Hydrates from Llquld Hydrocerbon-Watw Systems. Ph.D. Thesl8, UnivsrclHy of Mlchlgan, Ann Arbor, MI, 1974.

m.

s49

-

Schnelder, 0. R.; Farrar, J. US. Department of the Interlor, Research and Development No. 2 9 2 January 1968; p 37. R O W , 0. S.; Barduhn, A. J. &s&Wbi 1089, 6 , 57. Sloan, E. D., Jr. C%LtWere Hydvte.8 of Nnhrel@ses; Marcel Dekkw, Inc.: New York, 1990 p 641. (13) Fbyfd, F.; Sloen, E. D. Redlctkn of Nahral Qas Hydrate Dbmciatbn Enthalpks. Fkmwd@s of F h t I n m t h m a l oflbhae and pder EnCon-, 447 Ednkrgh, A w t 11-16, 1991. (14) van der Waals, J. H.; Plalteeuw, J. C. Clathrate sdutkns. A&. chem.P n p . 1050, 2 , 1. Recehred for review December 26, 1991. Revbed Aprll6, 1992. Accepted Aprll26, 1992. We are grateful for partiel suppart of thb work by the World Bank Development RoJect wlth the Instltute of Technology Bandq, Indona Sla

.

Viscosity of Liquid Toluene at Temperatures from 25 to 150 at Pressures up to 30 MPa

O C

and

Allrod H. Krall and Jan V. Songers'

Insmute for phvslcel S c h c e end Technology, Unhwsily of Maryland, college Pa&,

Maryland 20742

Joseph Kertln Dlvlskm of Enginwlng, Brown Unhersily, Rovkience, Rhode Island 02972 Now masuromonts are presented for tho vkcodty of lkpki tduone at twnporatures from 25 to 150 OC and at pressures up to 90 MPa. The meawretnents were obtahd wHh an osclllatlng-dkk vlacometer and have an ostlmated accuracy of h0.5% A comparkon wtlh data r w e d by othor rowarcham k Included, and an equation tor tho vkcodty of llqdd tduono as a function of temperature and d m d i y k given. Tduono was chosen k c a u w of Its uwtulnoss as a prlmary reference standard substance for thmnal conductlvtly of llqulds and Its potantla1 uw as a slmllar standard for viscosity.

.

Introduction I n this paper we present accurate measurements of the vlscoslty of llquid toluene as a functbn of both temperatwe and

pressure. The measurements were obtained with an osclilatingdisk viscometer. Our declsbn to measure the viscosity of liquid tduene was based on the foibwing Consideratbns. Liquid toluene has been recommended as a primary reference standard for the thermal conductivity of liquids ( 7 -3),and the question arlses whether liquid toluene would also be a suitable reference standard for the viscosity of ilquMs ( 4 , 5). Accordingly, the Subcommittee of IUPAC Commission 1.2 on the Transport Properties of Fluids has recommended that the viscosltyof liquldtduene be In different laboratories with a variety of experimental techniques. Osciilatingdisk viscometers have proven to be reliable insbunentsto meesuethevlscosltyofgasesand, morerecently, of liquids (6, 7 ) . The reason Is that considerable theoretical informatlon Is available concerning the working equatlons for this method ( 8 , 9). The theory of the working equations has been reconsidered by Kestin and Shankland ( 70, 7 7) and by Nleuwoudt et ai. ( 7 , 72)with the purpose of lnvestlgating how the i"t can be used for m e a m the vlscoslty of liquids both on an absolute and on a reiattve basis. Since the viscosity of ilquld toluene near room temperatwe has been measured by 0021-95~a1921i~37-0349$03.0010

a number of researchers, it was considered useful to make addknai measurements by the oscillatlngdisk method and 80 to provide reliable values for b use as a calibration substance for the viscosity of liquids. FvthemKne, toluene is a frequently used solvent, and reliable information about its physical prop e-, lnckrding vlscoeity, is also of dlrect technical importance. Experhnantal Mothod

The viscosity was measured with an osclilatlngdisk viscometer developed by Kestin and co-workers at Brown University (73, 74) and subsequently transferred to the University of Maryland (75). The viscometer is shown in Flgure 1. Nearly all the wetted parts are made of Hastelby C276 for resistance to conoskn. The viscometer body conskts of two large pleces. The upper piece 7 defines the cavity which is filled with the liquid to be investigated. The cavity volume Is about 1.3 dm3. The lower plece 3 carries a Bridgman window, 2, a filling port, 19, and a mukijunction thermocouple probe, 16. The two pleces are held together by a stainless steel cap, 8, which threads onto the lower plece and bears on the upper piece v k ten bob, 12, and a pressure ring, 13. The two pleces of the viscometer are sealed against an Internal pressure up to 30 MPa by means of a Viton O-ring, 17, or, for use at temperatures above 200 OC, by a metal k i n g , 20. The lower piece 3 rests on a self-centering roller bearing, 1. An osclilatbn of the disk is Initiated by manually twnlng the viscometer through a small angle, waltlng approximately one half-perbd, and then returning it to the orlglnai position. The oscillating disk 5 with a radius of 33.972 mm, a helght of 3.21 mm, and a mass of 102.1 g b suspended between two fixed horlzontai plates, 4 and 6. The gap between the dlsk and each plate is 2.249 mm. The purpose of the presence of the flxed plates is to suppress any possible convective fbw of the liquid near the dlsk. The oscillating disk is mountedto a torsion wire, 9, vla a cyllndricai chuck with a radlus of 2.79 mm and a height of 26.33 mm. The stress-reileved torsion wire has a length of 256 mm and a diameter of 0.13 mm; it Is made of a 0 1992 American Chemical Society

8

my---

pressurized with the aid of a hand pump, and the pressure is measured with a Bourdon gauge with an accuracy of 0.1 MPa. The viscometer is heated by an electric furnace which consists of an insulated resistance wire wrapped on an aluminum core in bifilar windings. The i n s b diameters of the me match the outside diameters of the viscometer top and the stainlesssteel cap. The power to the two zones is separately controlled by feedback controllers responding to thermocouple sensors in the windings of the fumace. The twolzone system allows a small positive temperature gradient to be imposed as a further safeguard against convection. The viscometer is designed for operation at temperatures up to 300 OC. However, in measuring the viscosity of toluene, we did not go beyond 150 OC. The problem is that the elastic properties of the torsion wire begin to vary appreciably at temperatures beyond 200 OC, causing some bss of accuracy at the higher temperatures ( 74, 76). Temperatures inside the viscometer are measured with copper-constantan thermocouples. Probe 21 with thermocouple junctions spaced over 30 cm allows measurement of the temperature t, of the suspension wire, upon which the period Toand the damping rate A, in vacuo depend. Probe 16 contains four junctions, spaced over 1.8 cm, which monitor the temperatwe distributionaround the disk. The third junction from the bottom is at the level of the disk, and tts reading is taken as the reference temperature t of the viscosity measurement. A digital voltmeter allows measurement of the temperature to be made with a precision of f0.03 OC. The temperature measurementsare accurate to f0.1 OC near room temperature and to f0.6 OC at 150 OC. After an initial deflection the disk suspended in the liquid executes a damped harmonic motion, so that the angular disas a function of the time T satisfiesthe equation placement a(~) (7- 72)

= a0exp(-2?rA~/T) sin (2?rT/ T)

CY(T)

Flgw 1. Osclllating-disk viscometer.

Table I. Characteristicsof the Oscillatina System 33.972 mm radius of disk R height of disk 2h 3.210 mm 2.249 mm spacing between disk and plates b 8.7941 g cm-3 density of disk (p 59093 g mm2 moment of inertia I length of torsion wire 256 mm diameter of torsion wire 0.13 mm period Toin vacuo at 25 O C 16.7894 s 3.4 x 104 damping rate A,, in vacuo at 25 O C 92% Pt-8% W alloy, since this material has a small internal damping and gives a high reproducibility of the rest position ( 76). The wire support and adjustment assembly 11 is carried by an alumina column, 10, which is open over most of its height. This alumina column is attached at its bottom to the suspension mounting plate 14 which is in tum attached to the cylindrical holder 15 carrying the fixed plates 4 and 6. The lengths and thermakxpansion coefficients of the column, the wire, and holder are such that the disk remains centered between the plates 4 and 6 over the working temperature range of the instrument (7). The column and holder are constructed in such a way that, with the upper piece 7 of the viscometer removed, the disk and wke are readily accessible. This permits checking and adjusting of the true running of the disk, the coincidence of the bearing axis with the suspension wire, and the disk's podtion between the two fixed plates. The characteristic properties of the oscillating system are summarlzed in Table I. The Viscometer is filled by albwing liquid to flow through the bwer port 19 from an elevated reservoir by gravity. A port at the top of the viscometer allows air to escape. The surfaces of the holder and column have been designed so as to avoid trapping air bubbles above the rising liquid. The system is

(1)

where a, is the amplitude, Tthe period, and A the damping rate of the oscillations. The period Tand the damping rate A are measured optically. A mirror, 18, is mounted to the oscillating disk through a cylindrical stem with a radius of 1.17 mm and a height of 101.6 mm. Laser light reflected from the mirror sweeps past two detectors, one located near the rest position of the oscillating beam and the other at an arbitrary, but fixed, offset angle. The period T and the damping rate A are then obtained from the time intervals registered by the detectors as described by Kestin and Shankland (77). The period Toof the oscillatingsystem in vacuo was deduced from the period T, measured in argon gas with a procedure described by Nieuwoudt et al. (72). I t is represented by T,/s

= 16.7894 + 2.115

-

X 103(t,/oC 25) 7.6 X lO-'(t,/OC

+

- 25p (2)

where t, is the temperature of the torsion wire. The period To is sb.ictly speaking not the period in vacuo, butthe period in the absence of the liquid. The effect of an elevated pressure on the oscillating system is primarily through the changed properties of the liquid, but there is also a small effect on the oscillating system directly resulting in a small dependence of T on the pressure P, such that T;l(t3To/t3PX = -9.2 X 10' MPa-'. This pressure correction is not important in measuring the viscosity, but it becomes relevant when one tries to use the oscillating-disk method for measuring liquid densities as well (75). The damping rate CI, may be similarly extracted from measurements in a gas; it has been previously determined by Nott as (78) with do= 3.668 X

lo",

dl = -1.146 X

lo",

d2 = 5.633 X

J m l of Chemical and Engineering Data, Vd. 37,No. 3,

10-l1,and d 3 = 1.612 X lo-,’. Further details concerning the 0scUlatlngdlsk instrument and its operation are contained in the Ph.D. thesis of Krali (79).

and coth (bs1’2/6) With

Worklng Equatlon

sin (Bb/6x)

The damped harmonic motkn described by eq 1 is governed by the equation

(s

+ AJ2 + 1 + D ( s ) = 0

K1(6)= cosh (2xb/6)

K2(8)= cosh (2xb/6)

s = (-A

+ i)To/T

(5)

where D(s) is the scaled Laplace transform of the torque exerted by the iiquld on the disk, while Toand 4 are the period and damping rate of the oscillating motion In vacuo (7- 72). The torque function D ( s ) is related to the viscosity 7 and the density p of the iiquM through the boundary layer thickness, defined as

6 = (7r,/2~p)l/~

- A#

= Im D ( s )

2(AO - Ao)8 = 81P6[HlK2 + H2Kl

With

e=Top An exact expressionfor the torque function D ( s )of the actual oscillatingdisk assembly is not available. However, for a disk with radius R, hdght 2h and density p , oscillating between two fixed horizontal plates located at equal distances 6 above and below the disk, a suitable approximate expression is (72, 75)

D ( s ) = @6/ph)[As3”

+ B s ( 6 / R ) + C S ~ ’ ~ ( ~ / R(9)) ~ ]

With

= coth ( 6 ~ ‘ / ~ / 6+) 4h/R

(10)

= ( 1 6 / 3 ~ ) ( 4 ~ / 2 7 ”-~1) + 6h/R

(11)

A B

+ 3h/2R

C = 17/9

where g, (i = 1-4) are instrument constants defined by g,

=x+&

(12)

(13)

Wlth

x

= ((e/2)[(A2 y=

+ 1)2-

e/&

(23)

= 4h/R

(24)

= B/R

(25)

= C/2R2

(26)

g, 93

g,

I n eq 22 we have introduced an unknown function E(6) to account for any deviations due to the theoreticalapproximations Introduced in the derivation of the working equation. The working equation (eq 22) reduces to the working equation adopted by Kestin and Shankland (70), tf the fkst terms in eqs 11 and 12 are taken to be zero. Inclusion of these terms has the advantage that the working equation (eq 22) with E(6) = 0 can be USBd to measure the viscosity of ilqulds on an absolute basis without calibration. As demonstrated in the next sectbn, such a procedure yields liqukl vlscoslties with accuracies at the percentage level. To obtain a hlgher accuracy, we retain the function E(6) and determine it by calibrating wlth a reference liquid. The calibration functbn E(6) in the working equation (eq 22) is related to the complex calibration function e(s J)introduced by Nieuwoudt et ai. (72) through E ( 6 ) = (h / R ) l m [ s ~ / ~ E ( I , ~ ) ] .

(14) (15)

The function E(6) in the working equation (eq 22) was determined by calibrating wlth liquid water. For thls purpose the viscometer was filled with distilled water degassed by vigorous boiling. The period and damping rate were measured at temperatures from 22 to 156 O C , with the temperature Intervals chosen so as to obtain roughly equal Intervals in the boundary layer thickness 6. Ail measurements were made at a slightly elevated pressure of 1.5 MPa so as to be well above the saturatbn pressure at ail temperaturesand to avoM the formation of vapor bubbles around the disk. The density ph,, and the viscosity qHp of the water were calculated from the equations recommended by Kestin et ai. (27). The values thus obtained for the function E(6) in eq 22 are shown in Figure 2 as a function of the boundary layer thickness. The measurements correspond to a range from 6 = 1.60 mm at room temperature to 6 = 0.72 mm at the highest temperature. The function E(6) can be represented by a cubic polynomial of the form E(6) = e o

With

= (ijh)-l = aR4/Z

Callbratlon

Equatbn 10 and the second terms of eqs 11 and 12 represent the torque function of a disk between two fixed plates without edge effects (20),while the first terms of eqs 11 and 12 represent an expansbn in terms of 6/R to account for edge effects in the approximation that the effect of the presence of the fixed plates on these edge effects can be neglected (9). To determine Im D(s), it is convenient to write (70) s1/2

+ 92Hl + 8386 + 84(8/x)62 + Wl (22)

(6)

(7)

- cos (&/ax)

We then obtain from eq 5 the final form of the working equation

I f the density p of the liquid is known, the viscosity 7 of the ilquld is deduced from the observed damping rate A and perkd Tthrough the imaginary part of eq 4: 2(Ae

- cos (&/6x)

sinh (2xb/6)

(4)

Wlth

and

= K2(6) - I Kl(6)

+ e,@/”)

+ e2(6/mm)’ + e3(6/mm)3

(27)

with e, = -0.075 725, e = 0.229 537, e 2 = -0.205 340, and e3 = 0.055056. Equation 27 reproduces the observed caiibration values of E(6) wlth a standard deviation of 0.00036 which correspondsto an uncertalnty In the measured viecoslty of only 0.06%.

952 Jownal of Chemlcal and Englineerlng Data, Vd. 37,No. 3, 19Q2 700

D v

w

-

a"3

0.00

600 500

.400

c

300 200 0.8

1.2

750

1.6

6 "I Flguv 2. CaRxatkn functkn €(a) as a function of the boundery layer thickness 6. +2

u 850

800

900

p I kg m-'

w 4. Measured vlscosltles of Uquld tohme plotted as a function of denstty: 0, 24 O C ; 0 , 3 0 O C ; V,41 O C ; V,49 'C 0 , 6 0 O C . , 77 'C A, 98 O C ; A* 121 0 , 152 'c. F

I

I

Intematknai Practical Temperature Scale of 1968 (IWS-68). For the purpose of thls paper ail temperatures have been converted to the new Intemathai Temperetve Scale of 1990 (ITS-90). For this purpose we represented the dlfferences between the temperatures T, and Tw on IWS-68 and ITS90 by a polynomial of the form (Tea

(b)

0

50

100

150

t 1°C Fbure 9. Comparison of the vlscodty q measuredfor llquid water as a functlon of temperature with the reference viscosity qw from the

literature,(a) when the viscometer Is used In the absolute mode and (b) when the viscometer Is used In the relathre mode. The actual accuracy of the measured vlscositbs will depend on the accuracy of the values of the viscoSrtieg as given by the equation of Watson et ai. (22) used in this paper. At t e m peratuw between 30 and 115 O C this equatlon Mers r fo m an alternate equation proposed by Kestln et ai. (23)by less than -0.2%, but at 150 O C It differs by +0.4%. The estimated accuracy of 10.4% at 150 O C causes some uncertainty in the curvature of E(6) observed at the smaller boundary layer thickness values. I t Is interestingto check to what extent our vlscometer can yleld viscosltles of IlquMs on an absolute basis. When the viscometer is used in an absolute mode, the viscosity 7 Is deduced ft" the measwed damping rate and perlod through the worklng equatbn (sq 22) with E(6) = 0. The absolute vlscoslties thus obtained are compared with the reference viscosbs in Figure 3a. We conclude that the Instrument yields absolute vlscositles of ilquid water with an accuracy of about f 1% . When the viscometer Is used In the relathre mode, the vlscosky 7 Is deduced from the measured damping rate and period through the worklng equatbn (eq 22) with E(6) represented by eq 27. The retetlve vlscositles thus obtained are compared wtth the reference vlscositles In Figure 3b. The relative viscositles are reproduced within 10.2% which Is within the estimated accuracy of the reference viscosities as discussed above. Experhentrl V ~ o r l t l e olor Toluene Hlghgwity liquid chromatography (HPLC) grade toluene with a stated pwlty of 99.9 % as suppiled by Aidrich Chemical was used forthe experiments. The vkoslty was determinedat nine temperaturesfrom 24 to 151 O C and at seven pressuresfrom atmospheric plesSW8 Up to 30 m. The mSaSUem8tlb W W 8 obtained and analyzed In terms of temperatures based on the

- Tg,)/K

+

= 8,(Tee/K - 273.16) 82(T@/K - 273.16)2 83(T,/K

+

- 273.16)3 (28)

with a, = 2.361 26 X lo4, 8 2 = 5.50424 X lo-', and 8 3 = -3.61922 X lo-'. Equatlon 28 is based on a table presented by McOleshan (24) and is valid in the range 270 K I TI 440 K. (It should be noted that in thls section the symbol Tdenotes temperatures In Keivln and not period as in the previous sections.) I n order to deduce the viscoSny from the observed damping rate and perlod through the worklng equation (eq 22), we need to know the denslty p of liquid toluene. For thls purpose we adopted an equation proposed by Kashlwagi et ai. (25):

( [

p/(kg m-3) = p , ( f ) 1 - C in

+ P/MPa D ( T ) + 0.1

D(T)

1)

(29)

with C = 0.09143 and W e the functions po( T ) and D(T) are polynomial functbns of the temperature T . This equation Is based on experimental densities at temperatures from 0 to 100 O C and pressures up to 200 MPa. Here we have extrapolated It for use at temperatures up to 150 O C . After conversion of the temperatures to ITS-90, the functions po( T ) and D(T) can be represented as p0(T) = 1103.15 - 0.68109(T/K)

- 0.4228 X

103(T/K)2 (29a)

and D ( T ) = 440.60

- 1.6047(T/K) + 1.540 X

10"(T/K)2 (29b)

The experimental results obtained for the viscosity of toluene are presented in Table 11; Flgure 4, showing the vlscosity as a function of denslty, gives an overview of the data set. Errors In our meesurements of the viscosky of liquid toluene may be ascribed to three sources: errors In the reference values of the viscosity of water used in the calibration of the vlscometer, Instrumentaldrltts or lack of long-term reproduclbiiity, and systemetlc errors of such a nature as not to be removed by the calibration. The range of boundary layer thickness covered in the toluene measurements, 0.84 mm I 6 5 1.45 mm, corresponds In water to temperatures In the range from 30 to 115 O C . The accuracy of the reference values of the viscosity of ilquid water may be considered to be f0.3% on this range (22, 23). We tested the viscometer far long-term stability by performlng a series of check measure-

Journal of Chem/cal and Engl-

24.52 29.09

862.9 858.7

559.5 529.1

41.14 48.64

24.56 29.35 41.17

866.7 862.3 851.5

581.8 548.5 480.6

48.65 59.40 76.59

24.31 29.44 41.10

870.7 866.1 855.7

605.9 569.3 500.7

48.62 59.34 76.75

24.39 28.64 41.10

874.2 870.5 859.6

628.3 597.3 520.4

48.63 59.37 76.92

24.40 28.72 41.12

877.7 874.0 863.4

651.7 619.4 539.7

48.02 59.62 77.04

24.46 28.99 41.23

881.0 877.2 866.9

674.7 640.5 558.6

48.71 59.80 77.16

24.48 29.25 42.29

884.2 880.3 869.6

700.8 661.2 571.6

48.91 59.89 77.39

847.3 840.2

Data, Vd. 37,No. 3, 1992 35s

462.3 427.3

59.28 76.54

830.0 813.3

384.5 328.1

444.9 399.9 342.2

98.22 121.02 151.28

797.9 776.0 746.4

286.1 240.5 194.4

462.8 416.7 356.1

98.32 121.21 151.52

803.6 782.4 754.0

298.4 251.7 204.9

480.8 432.5 369.7

98.83 121.51 152.39

808.5 788.3 760.5

309.7 262.6 214.4

501.7 448.2 383.8

98.53 121.65 152.53

813.9 793.9 767.0

323.1 273.6 224.2

517.0 464.2 398.0

98.79 121.72 152.68

818.5 799.3 773.1

335.2 284.7 234.0

534.0 480.3 411.4

99.00 122.09 152.86

823.0 804.1 778.8

346.8 294.8 243.2

P = 5.0 MPa 844.6 834.6 818.4

P = 10.0 MPa 848.9 839.2 823.3

P = 15.0 MPa 853.0 843.5 828.0

P = 20.0 MPa 857.4 847.5 832.4

P = 25.0 MPa 860.6 851.3 836.7

P = 30.0 MPa 864.1 855.1 840.7

ments of the viscosity of water after the completion of the measurements of toluene. Over the range of temperatures from 23 to 153 O C the viscosities calculated from eq 22 with E(6) given by eq 27 agreed with the reference viscosities to within f0.3%. Apart from the inaccuracy of the working equation which is corrected by the calibration function E@),the largest source of systematic error in our measurement is in the determination of the temperature. Like ail systematic errors, this one is partly removed by the relative mode of the measur-; however, because the condltlon of equal boundary layer thickness occurs In water and toluene at different temperatures and because water and toluene differ in their sensltMty to temperature, errm in the value assigned to the temperature have effects whlch are not entirely removed by a calibration function E(6) that depends on the boundary layer thickness abne. From consideration of the accuracy of the temperature measurements and their effects on the viscosity measurements, we judge^ that this source of error contributes an uncertainty of f0.2%. These three sources imply a total unwrtahty of f0.5%. Fvther details about the measuements can be found in the Ph.D. thesis of Kraii (19). I n order to represent the experimental viscosities as a functbn of temperature and density, we have adopted an equation previously used by Nieuwoudt et ai. for the viscosity of liquid isobutane (26): tl@',T')/bPa

8)

=

ho

h

-

With

p' = p / p c

T' = T/T,

where pc 290.2 kg m3 and T, = 593.91 K are the critical density and critical temperature of toluene (27). The coeffC c h t s h,in eq 30 are given in Table 111. The deviations of the measured viscosities from those calculated with eq 30 are shown in Figwe 5. The equation reproduces the experlmentai

Table 111. Coefficients in Equation 30 for the Viscosity of Liquid Toluene

12.977182 0.036 136 0.061266

h0

hl h3

f

c

+OS

0.271276 0.048892

h4 h6

I

- 1u 8

-0.5

&

-1.0

0

50

100

150

t/OC

Fl@"oS. Relatlvedeviatbmof themeesuedviscosltyfromthe~ qalo calculated with the coneletlng equation (eq 30): 0 , O . l We; V, 5.0 m;0, 10.0 Mpe; A, 15.0 MPa; 0 , 20.0 MPa; 0, 25.0 Mpa; V, 30.0 MPa.

viscosities with a standard deviation of 0.2%. Comparkon wfth Other Data

As a parallel project the viscosity of liquid toluene at atmospheric pressure has been measured in our laboratory by Gonpives et ai. with a high-precision Ubbebhde viscometer at temperatures from 25 to 75 O C . The difference between our new viscosity measurements obtained with the osciilatingdlsk method and those obtained by Gonplves et ai. varies from +0.3% at 25 O C to -0.3% at 75 OC, which is within the accuracy of f0.3% claimed by Gonplves et ai., but which r a veais a somewhat different trend of the temperature dependence of 7 . An extensive review of the viscosity of toluene at normal pressure has recently been presentedby Laesecke et ai. (28). An intercomparison of the various values reported for the vis-

354 Journal of Chemical and Englineerlng Data, Vol. 37,No. 3, 1992 +1.2

4.8

1

0

v

-0.4

8

-0.8

'

1

-t

-1.2

20

. ".

=*.

.r

_

n

v

v

0

40

0

0

80

60 t

O

0

D

1

pherlc pressure their results agree with those of Gkngaives et ai. to wlthln &OS% as shown in Figure 6; at elevated pressues they agree with our measurements within the clalmed accuracy of f4%. Assael et ai. (36)used a vibrating wke viscometer. Their data show small but systematic negative deviatlons from our results. However, except at the highest pressure of 30 MPa, these deviations are within the accuracy of f0.5% we claim for the correlating equatlon (eq 30). Acknowhdgment

100

1°C

Flgurr 6. Vlscoslty 7 of liquid toluene at atmospherlc pressure as a function of temperature, compared with the values qalC from an Arrhenkrs equatkn (eq 32)r e a m " d d by Qon@ahres et ai. (5)0, ref 5; 0, this work; V,ref 28; V, ref 32;0,ref 33; U, ref 31; A,ref 4; A ref 30; 0 ref 29; ref 35; X, ref 34.

We are Indebted to F. A. Gongaives, I. R. Shankland, D. Stewart, and W. A. Wakehem for valuable technical assistance and discussions. Q b r Y

b

E

. c

7F + ,

c

-2

-1.0

Z P R

: E d

L'

0

I

10

20

30

PIMPa

Fbun 7. Comparleon between the vlscoslty of liquid toluene at elevated pressures as reported In the literature with the values qDI1c calculated from our correlating yuatlon (eq 30): (a) falllngbody technlquo (95);0, 25 O C ; V, 50 C; 0,75 O C ; A, 100 O C ; (b) vibratinOcrysta1technlque (31);0, 25 O C ; V, 50 O C ; 0,75 O C ; (c) vlbratlng-wiretechnique (36);0, 30 O C ; V, 50 O C .

coslty of toluene at m i presswe (4,5,28-36) is presented in Figure 6; in this figure the data are shown as differences from an Arrhenlus equatlon: with A = 16.09 pPa s and B = 1055.4K as recommended by Qongahres et ai. (5). The measurements reported in refs 4,5, and 28-33 were obtained with capillary-flow viscometers. All but one of these are measurements relative to a reference Hqutd, usually water. Heydweller's result (a), an early absolute determination of the viscosity of water and of several other liqulds, is the exception. A small correction has been applied to the measurements of Heydweliier (29) and of Gelst and Cannon (30)so as to refer them to the current liqukl-viscosity benchmark, namely, 1002 pPa s for the viscoslty of water at 20 O C and atmospherlc pressure (3,37). Whenever possible, comparison with eq 32 has been made by convettlng the measured kinematic viscositles to dynamic viscositles wlth eq 29 for the density p. The highest accuracy, f0.2%, has been clalmed by Hammond et ai. (37)and Beuer and Meerlender ( 4 ) , but their results conslstently fail below those of Gkngahres et el. (5)by amounts sometimes exceeding 0.3%, but within 0.5%. Three determinationsof the viscosity of toluene at elevated pressures have been reported In the literature (34-36)with which our measurements can be compared; they are shown as devlatkns from ow correleting equatlon (eq 30) in Figure 7. Kashlwagl and Meklta measured the viscosity with a vibratlng pletoelectrlc crystal: they dld not report thelr orlglnai experimental data, but gave a formula for the pressure dependence of the viscodty on several isotherms from 25 to 75 O C which agrees with our results withln the clalmed accuracy of 2% . Dymond et ai. (35)used a faiilng-body viscometer. At atmoe-

S t

T TO TC T'

a

6 A A0

e

7 P Pc P'

P

distance between disk and plates caiibratlon function moment of lnertla pressure radius of dlsk Laplace-transform variable temperature, O C period or temperature, K period in vacuo critical temperature reduced temperature (= T I T c ) angular displacement boundary layer thlckness damping rate damping rate in vacuo inverse reduced period ( = T O I T ) viscosity density of liquid critical denslty reduced density (=PIP,) denslty of disk

Lltoraturo Clted Nieto de Castro, C. A.; Ll, S. F. Y.; Naga~J~lma, A.; Trengove, R. D.; Wakeham, W. A. J. Phys. Chem. Ref. Data 1988, 15, 1073. hmlres, M. L. V.; Vbka do8 Santos, F. Y.; Marddoer, U. V.; Nbto de Castro, C. A. Int. J. -ys. 1989, IO, 1005. Wakeham, W. A.; NagellMma, A.; Sengefs, J. V. W s " n t of Ihe T f a n m F m p T b s of F/&; Blackwell Sckntlfic: Oxford, 1091; pp 439-45 1. Bew, H.; Mwrbnder, Q. Rhed. Acta 1984, 23, 514. aOnwahnw, F. A.; Hemano, K.; Sengers, J. V.; Kestln, J. Int. J. W m y s . 1987, 8, 641. Kestln, J.; Wakeham, W. A. TmnspoH Rq" of Fhdds; Ho, C. Y., Ed.; CINDAS Data Serbs on Material Properties; Hemisphere: New York, 1088; Vol. 1-1. Nleuwoudt, J. C.; Shankland, I. R. Reference 3, pp 7-48. Keetln, J.; Newell, Q. F. ZAW 1967, 8, 433. AzpeMa, A. ci.; Newdl, Q. F. ZAW 1969, 10, 15. Kestin, J.; Shankland, I.R. .?AMP 1981, 32, 533. Kesnn, J.; shankland, I. R. J. Non-Eqtm. ~hevmodyn.i 9 8 i , 6 , 2 4 1 . Nieuwoudt, J. C.; Kestin, J.; Sengers, J. V. Pnyska A 1987, 142, 53. Kesfin, J.; Paul, R.; Shankland, I.R.; Khailfa, H. E. Ber. Bcnsen-(ks. Pnys. W m . 1980, 84, 1255. Kestln, J.; Imalshl, N.; Nott, S. H.; N&uwoudt, J. C.; Ssngers, J. V. p h f l k A 1965, 134, 38; 1988. 136, 617. Krall, A. H.; Nbuwoudt, J. C.; Sengers, J. V.; Kestln, J. F / M Pnese E q d b . 1987, 36, 207. Keetln, J.; Mossynakl, J. R. An Experimental Investigation of tho Internal Frlctkn of Thin PlaUnm Alloy W a s at Low Frequency. Brown University Report A7801/11; Brown Unlvemity: Provkbnce. RI, 1058. Kestln, J.; Shankland, I.R. Int. J. Thmnqhys. 1984, 5 , 3. Nott, S. H. Vlscc~sityof Light and Heavy Water and Their Mixtures. M.S. Thssls, Brown University, Providence, R I , 1084. Krall, A. H. 08cHhting-Body Vbcometry and b Applkatbn to pheseSeparatlng Uquld Mixtures. Ph.D. Thesk, University of Maryland, Cdlege Park, MD, 1002. Newell, Q. F. .?AM 1969, IO, 160. Kestln, J.; Seqers, J. V.; Kamgar-Pard, B.; Levelt Sengers, J. M. H. J . Phys. Chum. Ref. Deta 1904, 13, 175. Watson, J. T. R.; Beau, R. S.; Songera, J. V. J. Phys. them. Ref. Data 1980, P, 1255. Keetln. J.; Sokolov, M.; Wakeham, W. A. J. Phys. Chem. Ref. Deta 1978, 7, 041. McQhshan, M. L. J. Chem. 77"&n.1990, 22, 653.

J w n a l of Chemical and Englnmrlng Data, Vol. 37, No. 3, 1992 951 (25) Kahlwagi, H.; Hashlmoto, T.; Tanaka, Y.; Kubota, H.; Maklta. T. Int. J. llmmwphys. 1082, 3,201. (26) Nbuwoudt, J. C.; Le Neindre, 6.; Tufeu, R.; Sengers, J. V. J . Chem. Eng. Data 1987, 32, 1. (27) -In, R. D. J . phys. Chem. Ref. Data 1060, 18, 1565. (28) K a k , 6.;Laesecke, A.; Stelbrink, M. Int. J . Thennophys. 1091, 12, 289. (29) Heydweiller. A. Ann. Fhys. 1606, 59. 193. (30) W t , J. M.; Csnnon, M. R. Ind. Eng. Chem., Anal. Ed. 1946, 18, 611. (31) Hammond, L. W.; Howard, K. S.; McAliister, R. A. J. phys. Chem. 1058, 62, 637. (32) Dymnd, J. H.; Robertson, J. Int. J. hnmphys. 1085, 6, 21. (33) Byers, C. H.; Williams, D. H. J . Chem. Eng. Data 1087, 32, 344. (34) Kahiwagi, H.; Makita, T. rnt. J . "ophys. 1062, 3, 289.

(35) Dymond, J. H.; Awan, M. A.; Glen, N. F.; Isdeb, J. S. Int. J . 7h"yphvs. 1001, 12. 275. (36) Asiaei, M. J.; Papadaki, M.; Wakehem, W. A. Int. J . m y $ . 1901, 12, 449. (37) Swlndells, J. F.; Coe,J. R.; W e y , T. 6. J . Res. Net/. Bu. Stand. (U.S.) 1052, 48, 1. Received for revkw January 9, 1992. Accepted Aprli 8, 1992. This research has been pursued in consubtion wlth the Subcommlttw 1.2 on the Transport Ropedes of Flulds of IUPAC Commission 1.2 and is part of the research program of the Center for Envkonmentai Energy Engineering at the Unhrerslty of Maryland. The research was supported by the Pbburgh Energy Technology Center of the U.S. Department of Energy under @ant DEFG22-85PC80505.